Dan Hardy Memorial: Great St Mary’s, Cambridge, 2 February 2008

Young, Frances M.

doi:10.1177/0040571x12456666

Cited by 2 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The MAT was trained using the zrst [20], a python wrapper for the HTK toolkit [21] and srilm [22] that we developed for training unsupervised HMMs with varying model granularity. The LDA model we used in the Mutual Reinforcement was trained by MAL-LET [23].…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

An iterative deep learning framework for unsupervised discovery of speech features and linguistic units with applications on spoken term detection

Chung

Tsai

et al. 2015

2015 IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU)

View full text Add to dashboard Cite

In this work we aim to discover high quality speech features and linguistic units directly from unlabeled speech data in a zero resource scenario. The results are evaluated using the metrics and corpora proposed in the Zero Resource Speech Challenge organized at Interspeech 2015. A Multi-layered Acoustic Tokenizer (MAT) was proposed for automatic discovery of multiple sets of acoustic tokens from the given corpus. Each acoustic token set is specified by a set of hyperparameters that describe the model configuration. These sets of acoustic tokens carry different characteristics fof the given corpus and the language behind, thus can be mutually reinforced. The multiple sets of token labels are then used as the targets of a Multi-target Deep Neural Network (MDNN) trained on low-level acoustic features. Bottleneck features extracted from the MDNN are then used as the feedback input to the MAT and the MDNN itself in the next iteration. We call this iterative deep learning framework the Multi-layered Acoustic Tokenizing Deep Neural Network (MAT-DNN), which generates both high quality speech features for the Track 1 of the Challenge and acoustic tokens for the Track 2 of the Challenge. In addition, we performed extra experiments on the same corpora on the application of query-by-example spoken term detection. The experimental results showed the iterative deep learning framework of MAT-DNN improved the detection performance due to better underlying speech features and acoustic tokens.

show abstract

Section: Methodsmentioning

confidence: 99%

“…The LDA model we used in the Mutual Reinforcement was trained by MAL-LET [23]. The MFCC were extracted using the HTK toolkit [21]. The i-vectors were extracted using Kaldi [24].…”

Section: Methodsmentioning

confidence: 99%

An iterative deep learning framework for unsupervised discovery of speech features and linguistic units with applications on spoken term detection

Chung

Tsai

et al. 2015

2015 IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU)

View full text Add to dashboard Cite

show abstract

“…The two corpora used in the Zero Resource Speech Challenge were used here for easier comparison of results: the Buckeye corpus [37] (14137 utterances) in English and the NCHLT Xitsonga Speech corpus (4058 utterances) in Tsonga. The MAT was trained using the zrst [38], a python wrapper for the HTK toolkit [39] and srilm [40] for training unsupervised HMMs with varying model granularity. The ivectors were extracted using Kaldi [41].…”

Section: A Setup Of Mat-dnnmentioning

confidence: 99%

Unsupervised Iterative Deep Learning of Speech Features and Acoustic Tokens with Applications to Spoken Term Detection

Chung

Tsai

Liu

et al. 2017

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Abstract-In this paper we aim to automatically discover high quality frame-level speech features and acoustic tokens directly from unlabeled speech data. A Multi-granular Acoustic Tokenizer (MAT) was proposed for automatic discovery of multiple sets of acoustic tokens from the given corpus. Each acoustic token set is specified by a set of hyperparameters describing the model configuration. These different sets of acoustic tokens carry different characteristics for the given corpus and the language behind, thus can be mutually reinforced. The multiple sets of token labels are then used as the targets of a Multi-target Deep Neural Network (MDNN) trained on frame-level acoustic features. Bottleneck features extracted from the MDNN are then used as the feedback input to the MAT and the MDNN itself in the next iteration. The multi-granular acoustic token sets and the frame-level speech features can be iteratively optimized in the iterative deep learning framework. We call this framework the Multi-granular Acoustic Tokenizing Deep Neural Network (MAT-DNN). The results were evaluated using the metrics and corpora defined in the Zero Resource Speech Challenge organized at Interspeech 2015, and improved performance was obtained with a set of experiments of query-by-example spoken term detection on the same corpora. Visualization for the discovered tokens against the English phonemes was also shown.

show abstract

Dan Hardy Memorial: Great St Mary’s, Cambridge, 2 February 2008

Cited by 2 publications

References 0 publications

An iterative deep learning framework for unsupervised discovery of speech features and linguistic units with applications on spoken term detection

An iterative deep learning framework for unsupervised discovery of speech features and linguistic units with applications on spoken term detection

Unsupervised Iterative Deep Learning of Speech Features and Acoustic Tokens with Applications to Spoken Term Detection

Contact Info

Product

Resources

About